01-14-2008 12:33 AM - edited 03-15-2019 08:11 AM
Hi
Have a 3 server cluster running 5.1(3).
Appears the 2 subscribers are not replicating to their local databases. The environment works fine otherwise because the publisher database is referred to first and connectivity to the publisher database is fine from the subscribers.
The replication status of the subscribers is both 4. The CUCM reporting pages reports the local databases can't be contacted when running the DB state report.
When issuing "utils dbreplication repair all" on the publisher it says there are no other nodes in the cluster.
Looks like replication is not working at all, however the subscribers have full connectivity with the publisher database.
I suspect i need to rebuild the subscribers in order to rebuild the replication as the "util dbreplication rebuild" command does nothing.
I want to know if it might be the database access security password or something similar before i rebuild.
Any suggestions?
01-22-2008 07:22 AM
Could you verify if the Host and Lmhost files are up to date on all of the servers? So all of the servers hostname and IP addresses are present in these 2 files.
01-22-2008 03:46 PM
The CMhosts, rhosts and SQL hosts all appear consistent with names and IP's. Each server can ping the others via name or IP and there is no external DNS resolution in use.
I have rebuilt one of the subscribers but it didn't change anything in terms of the report output. The subscriber works fine, joins cluster and has 200+ registerd to it, but ther reporting page says its local database can't be reached.
Also noticed that i could disconnect the publisher and the 2 subscribers continued on fine as if their local databases were in fact working OK.
However the CM Database Status report in Cisco Unified Reporting still reports that that local databases of the 2 subscribers can't be reached. And the report is consistent no matter which host it is run on.
The cluster is running mixed mode security, if that makes any difference.
02-07-2008 12:55 PM
perhaps, this answer arrives a little later.
you should check the replicatiion status with utils dbreplication status
and check that there is a replication for each node. If there is not a replication for each node, you should do a reset all, utils dbreplication reset all but first stop replication in each node,utils dbreplication stop (subscribers first and the publisher the last).
It takes a long time.
Replication status must be 2
02-07-2008 03:12 PM
I tried all the CLI commands for stopping, reseting replication. In general they do nothing and if you do it on the Publisher it says that that there are no other nodes in the cluster.
All this despite if the publisher gets turned off the subsribers continue to operate as if they have OK local databases.
02-08-2008 10:48 PM
Hey buddy
Please let me know the exact steps and commands that you have followed for resetting the replication.
Tell me the exact topology and if you recently change any ipaddr or hostname/
Since you have 3 nodes I widely recommend you to open a TAC case with us to work directly in your server.
Or email me directly and I can give u some tips, the order of running the commands is really important
BTW I dont know of any util dbreplication rebuild you may be referring of clusterreset
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide